Resource scheduling method and device based on industrial topology network

Abstract

The embodiment of the invention discloses a resource scheduling method and device based on an industrial topology network. A specific embodiment of the method comprises the following steps: generating a corresponding operation information set, and generating a corresponding abnormal score sequence set based on the operation information set; calling a parallel computing node cluster of the topology construction server to generate an industrial topology network representing a product circulation relationship between production platforms; synchronously storing the abnormal score sequence set and the industrial topology network to a shared storage unit; through a parallel computing assembly line of a hardware accelerator, the following generation steps are executed: generating comprehensive abnormal information of a target production platform; generating industrial chain breakage probability information; generating a corresponding resource scheduling instruction; and controlling the resource control terminal to execute scheduling operation on the resources. According to the embodiment, accurate scheduling of industrial chain resources can be realized, time resources are saved, and production loss caused by supply chain interruption is reduced.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of computer technology, and more specifically to a resource scheduling method and apparatus based on industrial topology networks. Background Technology

[0002] Currently, in the field of supply chain risk early warning and emergency resource allocation, existing technologies mainly rely on isolated monitoring systems based on a single data source and scheduling decision-making mechanisms that depend on human experience.

[0003] However, when using the above-mentioned methods for resource scheduling, the following technical problems often arise: it is difficult to integrate multi-source data, and it is impossible to quickly build an industrial topology network that reflects the overall picture of multiple production platforms. This leads to a disconnect between early warning and resource scheduling, and it is impossible to quickly achieve an automated closed loop from risk perception to intervention execution, resulting in extended resource scheduling response time and production losses.

[0004] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure propose resource scheduling methods, apparatuses, electronic devices, and computer-readable media based on industrial topology networks to address one or more of the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide a resource scheduling based on an industrial topology network, including: in response to detecting that the inventory of a target product is lower than a target threshold, generating a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product; using an anomaly detection engine, generating a corresponding anomaly score sequence set based on the operational information set; based on the operational information set, calling a parallel computing node cluster of a topology construction server to generate an industrial topology network characterizing the product flow relationship between production platforms; synchronously storing the anomaly score sequence set and the industrial topology network in a shared storage unit for access by a hardware accelerator, wherein the shared storage unit is used for zero-copy data parallel access; and executing the following generation steps through the parallel computing pipeline of the hardware accelerator: generating comprehensive anomaly information of the target production platform based on the anomaly score sequence and the industrial topology network; generating industrial chain breakage probability information based on the comprehensive anomaly information, which is mapped to early warning level information; generating a corresponding resource scheduling instruction based on the early warning level information; and in response to receiving the resource scheduling instruction, controlling a resource control terminal to perform resource scheduling operations, wherein the scheduling operations include: storage of components related to the target product and transportation of the components.

[0008] Secondly, some embodiments of this disclosure provide a resource scheduling device based on an industrial topology network, comprising: a first generation unit configured to, in response to detecting that the inventory of a target product is lower than a target threshold, generate a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product; a second generation unit configured to, using an anomaly detection engine, generate a corresponding anomaly score sequence set based on the operational information set; a calling unit configured to, based on the operational information set, call a parallel computing node cluster of a topology construction server to generate an industrial topology network representing the product flow relationship between production platforms; and a storage unit configured to synchronously store the anomaly score sequence set and the industrial topology network in a shared storage unit. The system comprises: a shared storage unit for access by a hardware accelerator, wherein the shared storage unit is used for zero-copy data parallel access; an execution unit configured to perform the following generation steps through the parallel computing pipeline of the hardware accelerator: generating comprehensive anomaly information of the target production platform based on the anomaly score sequence and the industry topology network; generating supply chain disruption probability information based on the comprehensive anomaly information, which is mapped to early warning level information; a third generation unit configured to generate corresponding resource scheduling instructions based on the early warning level information; and a control unit configured to control a resource control terminal to perform resource scheduling operations in response to receiving the resource scheduling instructions, wherein the scheduling operations include: storage of components related to the target product and transportation of the components.

[0009] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0010] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0011] The above-described embodiments of this disclosure have the following beneficial effects: The resource scheduling method based on industrial topology networks, as described in some embodiments of this disclosure, achieves automated intervention (resource scheduling) for supply chain disruptions in production platforms, saving time and resources and reducing losses from supply chain interruptions. Specifically, the causes of supply chain disruption losses are: existing technologies suffer from severe data silos, making it difficult to correlate multi-source heterogeneous raw data, and lack risk propagation analysis from a supply chain perspective, leading to delayed risk identification, insufficient breakpoint prediction, and an inability to take timely intervention measures, resulting in supply chain disruptions. Based on this, the resource scheduling method based on industrial topology networks, as described in some embodiments of this disclosure, firstly, in response to detecting that the inventory of a target product is below a target threshold, generates a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the supply chain corresponding to the target product. The multi-source heterogeneous raw data is uniformly encoded into a standardized time-series vector, forming a continuous and measurable operational information set, providing a unified and comparable basic data representation for subsequent analysis. Then, using an anomaly detection engine, a corresponding anomaly score sequence set is generated based on the above operational information set. By generating the anomaly score sequence set, the degree of deviation of the production platform's state is accurately quantified, improving the sensitivity of early weak anomaly signal identification and providing an intuitive quantitative basis for risk assessment. Next, based on the aforementioned operational information set, the parallel computing node cluster of the topology construction server is invoked to generate an industry topology network representing the product flow relationships between production platforms. By constructing this industry topology network, supply relationships between production platforms can be efficiently mined, supporting real-time monitoring of large-scale production platforms and clearly presenting the industry chain structure. Secondly, the aforementioned anomaly score sequence set and industry topology network are synchronously stored in a shared storage unit for access by the hardware accelerator. This shared storage unit is used for zero-copy data parallel access. This eliminates the data copying overhead between the CPU and the hardware accelerator, ensuring data consistency and providing a low-latency, high-bandwidth data supply channel for subsequent hardware-accelerated computing. Thirdly, through the parallel computing pipeline of the hardware accelerator, the following generation steps are executed: Based on the aforementioned anomaly score sequence set and industry topology network, comprehensive anomaly information for the target production platform is generated. The parallel computing pipeline improves the efficiency of multi-source data fusion, integrates anomaly scores and topology relationships, and achieves a comprehensive assessment of production platform risks, avoiding the one-sidedness of isolated analysis. Based on the aforementioned comprehensive anomaly information, industry chain breakage probability information is generated and mapped to early warning level information. The risk of supply chain disruption is quantified by presenting the risk level intuitively through early warning levels, facilitating rapid response and addressing the delays caused by ambiguous risk levels. Next, based on the aforementioned early warning level information, corresponding resource scheduling instructions are generated. This achieves precise matching between anomalies and scheduling strategies, ensuring the targeted and effective nature of the scheduling instructions. Finally, in response to receiving the resource scheduling instructions, the resource control terminal executes resource scheduling operations, including: the storage and transportation of components related to the target product.Rapidly deploying component storage and transportation scheduling, timely replenishing production resources, reducing the probability of supply chain disruptions, ensuring the continuity of target product production, responding quickly to risks, and minimizing losses from supply chain disruptions. Attached Figure Description

[0012] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0013] Figure 1 This is a flowchart of some embodiments of the resource scheduling method based on industrial topology networks according to this disclosure; Figure 2 These are schematic diagrams illustrating the structure of some embodiments of the resource scheduling device based on an industrial topology network according to this disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0014] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0015] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0016] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0017] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0018] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0019] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] refer to Figure 1 The diagram illustrates a flow 100 of some embodiments of a resource scheduling method based on an industry topology network according to the present disclosure. This resource scheduling based on the industry topology network includes the following steps: Step 101: In response to detecting that the inventory of the target product is lower than the target threshold, a corresponding set of operational information is generated based on the multi-source heterogeneous raw data of multiple production platforms in the industrial chain corresponding to the target product.

[0021] In some embodiments, the executor of the resource scheduling method based on the industrial topology network described above (e.g., an electronic device) can be hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster composed of multiple servers or terminal devices, or as a single server or a single terminal device. When the computing device is software, it can be installed in the hardware devices listed above. It can be implemented as multiple software programs or software modules to provide distributed services, or as a single software program or software module. No specific limitations are made here.

[0022] In other embodiments, the aforementioned executing entity may, in response to detecting that the inventory of the target product is lower than a target threshold, generate a corresponding set of operational information based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product. The target product may be a specific product requiring production (including multiple components, and requiring supply from multiple production platforms), such as new energy vehicles or mobile phones. The target threshold may be a pre-set critical value for the inventory of the target product (e.g., the target threshold for new energy vehicles may be 200 vehicles). The multi-source heterogeneous raw data may be operational data from production platforms in different channels and formats. For example, multi-source heterogeneous raw data may include: business registration change records (text), recruitment frequency (numerical), and public opinion comments (text). The set of production platform behavior trajectory vector sequences may be a collection of quantified multi-dimensional state vectors arranged in chronological order from the multi-source heterogeneous raw data. The operational information may be a vector composed of multiple standardized operational dimension values ​​at a specific point in time, used to characterize the comprehensive state of the production platform at that specific point in time. For example, the above operational information could be [Business Registration Change: 0, Recruitment Index: 0.7, Logistics Index: 0.5, Invoice Strength: 0.8, Public Opinion Risk: 0.2]. For example, the operational information set could include: {Operational information of production platform A, Operational information of production platform B, Operational information of production platform C}, where each sequence is a multi-dimensional vector arranged by date.

[0023] In some optional implementations of certain embodiments, the execution entity may, in response to detecting that the inventory of the target product is lower than a target threshold, generate a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product. This may include the following steps: The first step is to perform the following steps on each of the multi-source heterogeneous raw data from the aforementioned multiple production platforms: Sub-step one involves performing heterogeneous data time-series isomorphism processing on the aforementioned multi-source heterogeneous raw data from the production platform to obtain isomorphic time-series data. This heterogeneous data time-series isomorphism can involve uniformly processing different types of data into a sequence format that can be aligned and compared on the time axis. The isomorphic time-series data can be time-series data obtained by converting the multi-source heterogeneous raw data from the production platform to a unified time granularity and format; for example, converting business registration changes (events) into 0 / 1 pulse sequences (daily dimension). In practice, firstly, based on the type of multi-source heterogeneous raw data from the production platform, event-type data (e.g., business registration changes) and continuous data (e.g., invoice amounts) are distinguished. Then, for event-type data, days with occurrences are marked as 1, and days without events are marked as 0, generating pulse sequences. For continuous data, daily aggregation (e.g., summation, counting) is performed to generate numerical sequences. Finally, all sequences are aligned according to a unified timestamp (e.g., daily), and missing days are marked to obtain isomorphic time-series data.

[0024] Sub-step two involves constructing a standard multidimensional trajectory vector based on the aforementioned homogenized time series data. This standard multidimensional trajectory vector can be a standardized vector of each dimension from multiple time points for the production platform. In practice, firstly, for each time point (e.g., daily), the values ​​of each dimension are extracted from the homogenized sequences of each production platform. Then, for each continuous dimension value, it is normalized using the recent historical extreme values ​​of the production platform (e.g., the past 90 days), scaling it to the [0, 1] interval. Finally, the dimensions of each production platform at each time point (0 / 1 for event-related data, normalized values ​​for continuous data) are combined into a fixed-dimensional vector.

[0025] Sub-step three involves maintaining the continuity of the aforementioned standard multidimensional trajectory vectors to generate corresponding operational information. In practice, firstly, it is determined whether there are any missing days in the daily vector sequence of each production platform. Then, for the missing day vectors of each production platform, linear interpolation is performed using the vectors of its preceding and following days to fill them in. Finally, it is determined whether the vector sequence of each production platform after filling is continuous in time (without discontinuities) and whether the first-order difference quotient can be calculated (approximately differentiable), thereby generating a complete set of production platform behavior trajectory vector sequences.

[0026] The second step is to integrate the various operational information into an operational information set.

[0027] Step 102: Using the anomaly detection engine, generate the corresponding anomaly score sequence set based on the runtime information set.

[0028] In some embodiments, the aforementioned execution entity can utilize an anomaly detection engine to generate a corresponding set of anomaly score sequences based on the aforementioned operational information set. The anomaly score sequences characterize the degree of deviation of the production platform's state. The anomaly detection engine can be a server used for automated analysis of the production platform's behavioral trajectory and identification of abnormal states. The anomaly detection engine can determine the degree of anomaly in the behavior of multiple production platforms by identifying their corresponding operational information sets (e.g., quantitative information on production, energy consumption, and orders), and generate a continuous numerical sequence. The anomaly score sequence set can be a combination of anomaly quantitative sequences from multiple production platforms. For example, the anomaly score sequence can be the daily anomaly score set of production platform A for the past 7 days. For example, the anomaly score sequence could be [0.1, 0.15, 0.3, 0.5, 0.8, 0.9, 0.95]. The degree of deviation of the production platform's state can be the difference between the production platform's current operating state and the normal baseline.

[0029] In some optional implementations of certain embodiments, the aforementioned execution entity may utilize an anomaly detection engine to generate a corresponding anomaly score sequence set based on the aforementioned runtime information set, which may include the following steps: The first step is to perform the following steps for each piece of operational information in the above operational information set: In the first sub-step, the first node of the anomaly detection engine is used to generate the trajectory curvature change sequence corresponding to the aforementioned operational information. This first node can be a node within the anomaly detection engine specifically responsible for calculating trajectory curvature (acceleration). The trajectory curvature change sequence can be a sequence characterizing the degree of change (acceleration) of the production platform's behavioral trajectory at each time point. For example, the trajectory curvature change sequence could be a production platform trajectory curvature sequence of [0.01, 0.02, 0.05, 0.12, 0.3, 0.6]. In practice, based on the operational information, the first node of the anomaly detection engine is invoked to determine the second derivative of the trajectory corresponding to the operational information. Finally, the derivative results are integrated to generate the trajectory curvature change sequence.

[0030] Sub-step two involves utilizing the second node of the anomaly detection engine to perform a dual-benchmark dynamic discrimination operation based on the aforementioned trajectory curvature change sequence, the preset historical normal trajectory benchmark of the production platform itself, and the industry average trajectory benchmark for the same period, to obtain dual-benchmark deviation data. The aforementioned second node can be a dedicated processing unit within the anomaly detection engine for performing dual-benchmark dynamic discrimination. The preset historical normal trajectory benchmark of the production platform itself can refer to the behavioral trajectory of the production platform during its past normal operating period as a comparison standard. For example, the preset historical normal trajectory benchmark of the production platform itself can be the average behavioral trajectory of production platform B over the past 12 months after excluding abnormal months. The aforementioned industry average trajectory benchmark for the same period can refer to the average behavioral trajectory of production platforms in the same industry during the same period. The aforementioned dual-benchmark deviation data can refer to the deviation values ​​(self-deviation and industry deviation) obtained by comparing the current trajectory of the production platform with its own historical benchmark and the industry benchmark, respectively. In practice, firstly, the trajectory curvature change sequence is compared with the production platform's own historical normal curvature benchmark using DTW (Dynamic Time Warping) to determine the shape difference distance. Then, it is compared with the industry average curvature benchmark for the same period using EMD (Earth Movement Distance) to determine the distribution difference. Finally, the two distance values ​​are normalized to obtain a deviation score to obtain dual-benchmark deviation data.

[0031] Sub-step three involves generating an anomaly score sequence based on the aforementioned dual-benchmark deviation data and adaptive weight parameters. The adaptive weight parameters are dynamically determined by the anomaly detection engine based on the completeness of historical data from the production platform and its historical warning accuracy. Specifically, the adaptive weight parameters can refer to a coefficient that dynamically adjusts the importance of the anomaly to its own deviation from the industry standard. The completeness of historical data from the production platform can refer to the proportion of valid data in the platform's historical data. The historical warning accuracy can be the proportion of past warning results that have been verified as correct. In practice, firstly, the anomaly detection engine determines the adaptive weight parameters (e.g., 0.7 (self-deviation) and 0.3 (industry deviation)) based on the completeness of historical data from the production platform (e.g., 92%) and the historical warning accuracy (e.g., 88%). Then, an anomaly score sequence is generated (e.g., according to "anomaly score = 0.7 × self-deviation + 0.3 × industry deviation").

[0032] Step 103: Based on the runtime information set, invoke the parallel computing node cluster of the topology construction server to generate an industry topology network that represents the product flow relationship between production platforms.

[0033] In some embodiments, the aforementioned execution entity can invoke a cluster of parallel computing nodes of a topology construction server based on the aforementioned runtime information set to generate an industry topology network representing the product flow relationships between production platforms. The topology construction server can have multiple parallel nodes, serving as a server for mining supply relationships between production platforms and constructing the network. The product flow relationships between production platforms can be upstream and downstream supply and demand relationships between production platforms, such as a directed cooperative relationship where component manufacturers supply parts to vehicle manufacturers. The aforementioned industry topology network can be a network structure where nodes represent production platforms and directed edges represent supply relationships.

[0034] The first step involves calling the first parallel computing node in the aforementioned parallel computing node cluster to determine the event-type trajectory similarity between every two production platforms in the aforementioned runtime information set, thereby generating an event synchronization matrix. The first parallel computing node can be the node in the topology building server that determines the event-type trajectory similarity. The event type refers to a discrete state where data occurs or does not occur at a specific point in time, typically represented by 0 (not occurred) or 1 (occurred). The event-type trajectory similarity can be the degree of synchronization between the event-type trajectories of the production platforms, for example, the synchronization matching degree between the capital increase of production platform A and the recruitment peak of production platform B. The event synchronization matrix can be a square matrix storing the event synchronization degrees between all pairs of production platforms. For example, the value in the i-th row and j-th column of the event synchronization matrix represents the event synchronization degree between production platform i and production platform j.

[0035] As an example, firstly, the first parallel computing node of the topology building server determines the event-type trajectories (e.g., business and public opinion pulse sequences) of each production platform. Then, for each pair of production platforms, the dynamic time-warped distance of its event sequences is determined within a allowed time offset (e.g., 0 to 7 days), and the distance is converted into a similarity score. Finally, the similarity results of each pair of production platforms are populated into a matrix to generate an event synchronization matrix.

[0036] The second step involves invoking the second parallel computing node in the aforementioned parallel computing node cluster to determine the numerical correlation between the continuous trajectories of every two production platforms in the aforementioned operational information set, thereby generating a numerical correlation matrix. This second parallel computing node can be a node in the topology building server that determines the correlation of continuous trajectories. The numerical correlation of these continuous trajectories can be the statistical correlation of the changing trends of continuous indicators (such as logistics volume) between the two production platforms. For example, the positive correlation between the daily logistics frequency sequences of the two production platforms. The numerical correlation matrix can be a square matrix storing the pairwise numerical correlations between all production platforms. For example, each element in the numerical correlation matrix is ​​a correlation coefficient (e.g., 0.9 indicates a strong positive correlation). In practice, firstly, the continuous trajectories (logistics frequency, recruitment index) of the aforementioned production platforms are extracted. Then, the second parallel computing node determines the Pearson correlation coefficient between the sequences of every two production platforms. Finally, the numerical correlation matrix is ​​generated.

[0037] The third step involves calling the third parallel computing node in the aforementioned parallel computing node cluster to determine the time-lag causal relationship between the operational information of each pair of production platforms in the aforementioned operational information set, in order to generate the optimal time-lag matrix. This third parallel computing node can be a node in the topology building server specifically dedicated to parallel computing time-lag causal relationships. The time-lag causal relationship can be the number of days and the degree of correlation between the changes of one production platform and the changes of another. For example, the time-lag causal relationship could be a change in the invoice amount of production platform A, followed by a similar change in the invoice amount of production platform B 3 days later, with production platform A leading production platform B by 3 days. The optimal time-lag matrix can be a square matrix storing the time-lag days corresponding to the maximum correlation between each production platform. The optimal time lag can refer to the number of days when the behavioral change of one production platform leads the change of another, and the correlation between the two is at its maximum. In practice, firstly, the third parallel computing node acquires the continuous trajectory of each production platform. Then, for each pair of production platforms, the sequence of one production platform is shifted relative to the sequence of the other by 1 to 7 days, the correlation coefficient under different time lags is determined, and the time-lag days with the maximum correlation are recorded. Finally, the optimal time delays for each production platform pair are filled into the matrix to obtain the optimal time delay matrix.

[0038] The fourth step involves weighted fusion of the aforementioned event synchronization matrix, numerical correlation matrix, and optimal time delay matrix to generate a comprehensive similarity matrix between production platforms. This comprehensive similarity matrix can be a holistic correlation matrix that integrates the event synchronization matrix, numerical correlation matrix, and optimal time delay matrix. In practice, firstly, weights are assigned to the three types of matrices (event synchronization matrix 0.3, numerical correlation matrix 0.4, optimal time delay matrix 0.3). Then, the event synchronization, numerical correlation, and optimal time delay matrices are summed according to their weights; finally, the comprehensive similarity matrix between production platforms is obtained.

[0039] The fifth step involves using the pre-acquired importance scores of the production platforms to weight and correct the overall similarity matrix between the production platforms, thereby generating a coupling strength matrix between them. The importance scores of the production platforms can be pre-calculated node importance indicators based on factors such as platform size, market share, and network centrality. For example, production platform A, with a high share of regional GDP and large tax payments, might have an importance score of 0.9 (out of 1). The coupling strength matrix between the production platforms can be a relationship strength matrix weighted by the importance scores, resulting in more reasonable edge weights. In practice, first, the importance scores of each production platform are obtained (e.g., production platform A has a score of 0.9, and production platform B has a score of 0.8). Then, the corresponding overall similarity is corrected using the product of the production platform importance scores. Finally, the coupling strength matrix between the production platforms is generated.

[0040] The sixth step involves filtering the coupling strength matrix between the aforementioned production platforms to generate an initial directed industry chain network. This initial directed industry chain network can be a basic network formed after filtering for a coupling strength threshold; for example, only directed edges between production platform pairs with a coupling strength greater than 0.5 are retained. In practice, first, a coupling strength threshold (e.g., 0.5) is set. Then, production platform pairs with coupling strengths greater than the threshold are filtered from the matrix. Finally, the initial directed industry chain network is constructed.

[0041] The seventh step involves noise filtering of the initial directed network of the industrial chain to generate an industrial topology network. In practice, firstly, the Louvain algorithm is applied to the initial directed network of the industrial chain to divide it into communities. Then, noisy edges (e.g., pseudo-association edges between production platform A and production platform B, which can be erroneous network connections based on data coincidence rather than real business relationships (e.g., two companies located in the same industrial park but with unrelated businesses, whose "event synchronization" due to simultaneously responding to regional power rationing policies leads to an edge that is misjudged by the algorithm as having a supply chain connection)) are removed to obtain the industrial topology network.

[0042] The above-described steps, as an inventive point of this disclosure, solve the technical problem of "the inability to efficiently and accurately automatically identify real supply relationships from massive, dynamic production platform behavior data, leading to a waste of computing and time resources." The reasons for this technical problem are as follows: existing technologies cannot automatically and efficiently mine accurate, time-series causal supply relationship networks from multi-source production platform behavior data, relying on manual experience or static correlations, which is insufficient to support large-scale, dynamic supply chain risk analysis. This invention utilizes a topology construction server to parallelly compute the similarity of multiple dimensions of operational information sets and fuse them into a high-precision supply chain network, achieving the automatic and efficient construction of a high-quality industrial topology network from operational information sets, saving significant time and computing resources.

[0043] Step 104: Synchronously store the abnormal score sequence set and the industry topology network to the shared storage unit for access by the hardware accelerator.

[0044] In some embodiments, the aforementioned execution entity may synchronously store the aforementioned abnormal score sequence set and the aforementioned industry topology network to a shared storage unit for access by a hardware accelerator. The shared storage unit is used for zero-copy data parallel access. The shared storage unit may be a hardware / software architecture that supports direct and seamless access to the same physical or virtual memory by the CPU and hardware accelerators (e.g., GPUs), eliminating data copying and achieving efficient data sharing. The aforementioned hardware accelerator may be a dedicated processor, such as a GPU or FPGA, specifically designed to accelerate specific computational tasks (e.g., graph computation, matrix operations), possessing massively parallel computing capabilities.

[0045] In addressing the technical challenges of the aforementioned background technologies, and considering the application scenario—the automated production process of automobiles under large-scale orders (orders with strict time delivery requirements), where the assembly of various components is highly collaborative and involves multiple production platforms—it is necessary to quickly locate production platforms with a high risk of disruption. This requires synchronizing massive amounts of abnormal production platform data and topology-related data to hardware accelerators within minutes to pre-emptively schedule necessary components. However, this often leads to the following technical problems: low heterogeneous data transmission efficiency and high hardware access addressing latency, resulting in delayed risk warning responses and consequently delayed resource scheduling responses, making it impossible to promptly mitigate losses from disruptions. Based on the following requirements for this application scenario—low data synchronization latency, zero hardware access redundancy, heterogeneous data classification and transmission, and high task execution concurrency—we have decided to adopt the following solution: In some optional implementations of certain embodiments, the execution entity may synchronously store the aforementioned abnormal score sequence set and the aforementioned industry topology network in a shared storage unit for access by the hardware accelerator, which may include the following steps: The first step is to allocate a corresponding target logical address space in the shared memory unit based on the aforementioned abnormal score sequence set and the data size descriptor of the aforementioned industry topology network. This target logical address space can be a pre-defined, contiguous virtual address range within the shared memory unit for a specific dataset, available for CPU writing and hardware accelerator reading. In practice, first, the data size descriptors of the abnormal score sequence set and the industry topology network are determined, and the required memory space size for each is obtained. Then, two contiguous and aligned logical address spaces are requested from the memory manager of the shared memory unit to serve as the target logical address space.

[0046] The second step is to establish a direct memory access channel based on the target logical address space. This memory access channel can be a hardware channel that allows external devices to bypass the CPU and directly perform high-speed data transfer with main memory, thus reducing CPU load. In practice, first, the physical address or I / O virtual address corresponding to the target logical address space is obtained. Then, the DMA channel parameters between the CPU and the shared memory unit controller are configured to establish the direct memory access channel.

[0047] The third step involves data sharding of the aforementioned abnormal score sequence set to generate a time-series data shard set. This time-series data shard set can be a collection of smaller, independent data blocks into which the abnormal score sequence set is divided according to the parallelism of the computational units, facilitating parallel processing. In practice, firstly, the optimal sharding granularity is determined based on the number of cores and cache size of the hardware accelerator. Then, the abnormal score sequence set is divided according to time windows. Finally, multiple evenly sized, internally contiguous data shards are generated to serve as the time-series data shard set.

[0048] The fourth step is to perform structural encoding on the aforementioned industry topology network to generate a topology data block set. This topology data block set can be a logical data block set (topology data block set) formed by encoding the graph structure data of the industry topology network into a format suitable for parallel hardware access (e.g., CSR).

[0049] The fifth step involves generating a transmission task flow based on the relationships and transmission priorities between the aforementioned time-series data shards and topology data block sets. These relationships can be computational dependencies between time-series data shards and topology data blocks, determining the order of data transmission and computation. For example, calculating the risk of production platform A depends on the anomaly score of its upstream supplier B; therefore, shards from B should be transmitted first. The transmission priorities can be a hierarchical ranking of data transmission tasks, ensuring critical data arrives first to initiate computation as early as possible. For example, anomaly shards from the core production platform and topology blocks from the network hub are given the highest priority. The concurrent transmission task flow can be an ordered set of data transmission tasks. First, the dependencies between shards and data blocks are analyzed (e.g., computation requires edge and node attributes). Then, tasks are prioritized (e.g., topology blocks first, followed by critical path production platform shards). Finally, a task list is generated, where tasks with no or low dependencies are marked for concurrent execution to generate the transmission task flow.

[0050] The sixth step involves executing the transmission tasks of the aforementioned transmission task flow through the direct memory access channel to write the aforementioned abnormal score sequence set and the aforementioned industry topology network into the aforementioned target logical address space, resulting in the target logical address space after data writing. This target logical address space after data writing can be a logical address space where data writing is complete and the content is ready. In practice, firstly, according to the transmission task flow, the source address (CPU memory) and destination address (target logical address space within the shared memory unit) of each task are obtained. Then, the DMA engine initiates multiple transmission tasks in parallel, bypassing the CPU, and reads data from the source address and writes it to the destination address through the established direct memory access channel. Finally, after all transmission tasks are completed, the target logical address space stores the complete and organized abnormal score fragment set and topology data block set.

[0051] Step 7: Based on the target logical address space and data type identifiers after the data is written, a global memory mapping table is generated. This global memory mapping table can be a global directory table recording the location, size, and type of all data blocks in shared memory; it serves as a "map" for the hardware accelerator to access the data. The data type identifiers are used to distinguish labels for different data structures (e.g., vectors, matrices, graphs) in shared memory, facilitating correct parsing by the hardware accelerator. In practice, first, after each data transfer is completed, the target address space of the written data is scanned. Then, according to the predefined "data type identifiers," the starting address, size, and type of each data block are recorded. Finally, the global memory mapping table is generated.

[0052] Step 8 involves configuring the aforementioned global memory mapping table into the hardware accelerator to establish a zero-copy access path for the target logical address space after the data is written. This zero-copy access path allows the hardware accelerator to access data in shared memory directly via pointers without CPU copying, resulting in extremely low latency. In practice, firstly, the generated global memory mapping table is sent to the hardware accelerator via a driver program or dedicated instructions. Then, the hardware accelerator kernel driver creates an internal pointer based on the global memory mapping table, enabling its computational units to directly "see" the corresponding address in shared memory. Finally, the accelerator's computational threads can directly read and write data through this pointer, completing the establishment of the zero-copy path.

[0053] The above-described steps, as an inventive point of this disclosure, solve the technical problem mentioned in the background art: "low efficiency of heterogeneous data transmission and high hardware access addressing latency, leading to delayed risk warning response, which in turn leads to delayed resource scheduling response and inability to avoid losses due to supply chain disruptions in a timely manner." The reasons for these technical problems are as follows: no classification and transmission strategy was designed for time-series / topological heterogeneous data, and there is a lack of direct addressing mechanism between hardware and storage, resulting in high data copy redundancy. This invention, through heterogeneous data classification processing, concurrent transmission, and global memory mapping table configuration, achieves low-latency synchronous storage and zero-copy hardware access for abnormal score sequence sets and industrial topology networks, saving time costs for data transmission and hardware addressing, and reducing production losses caused by supply chain disruptions during the production process.

[0054] Step 105: Execute the following generation steps through the parallel computing pipeline of the hardware accelerator: Step 1051: Generate comprehensive anomaly information for the target production platform based on the anomaly score sequence set and the industry topology network.

[0055] In some embodiments, the aforementioned execution entity can generate comprehensive anomaly information for the target production platform based on the aforementioned anomaly score sequence set and the aforementioned industry topology network. The target production platform can be a specific production platform requiring risk analysis and intervention scheduling. The comprehensive anomaly information can be a time-series quantified sequence integrating its own and industry chain-related risks. The aforementioned parallel pipeline can be a high-efficiency execution model in a hardware accelerator that decomposes computational tasks into multiple independent stages and allows different data to flow and be processed simultaneously at different stages to maximize hardware utilization. For example, to determine comprehensive anomaly information for 1000 production platforms, the pipeline can be designed as follows: Stage 1 (parallel): extracting the upstream and downstream neighbor sets for each production platform in parallel; Stage 2 (parallel): calculating the upstream / downstream aggregate value for each production platform in parallel; Stage 3 (parallel): fusing the three types of scores for each production platform in parallel. Thus, when the data from the first batch of production platforms enters Stage 2, the data from the second batch of production platforms can immediately begin the calculation in Stage 1, achieving pipelined parallel throughput.

[0056] In some optional implementations of certain embodiments, the execution entity generates comprehensive anomaly information of the target production platform based on the aforementioned anomaly score sequence set and the aforementioned industry topology network, which may include the following steps: The first step is to extract the upstream and downstream neighbor sets of the target production platform based on the aforementioned industry topology network. The upstream neighbor set can be the set of production platforms in the industry topology network that supply the target production platform. The downstream neighbor set can be the set of production platforms in the industry topology network that receive products from the target production platform. In practice, firstly, the target production platform (e.g., an electronics manufacturer) is located based on the industry topology network. Then, the industry topology network is traversed to determine the upstream supplying production platforms pointing to the target production platform and the downstream production platforms pointed to by the target production platform.

[0057] The second step involves determining the abnormal score information corresponding to the target production platform, the upstream neighbor set, and the downstream neighbor set within a preset time window, based on the aforementioned abnormal score sequence set. This yields the abnormal score information of the target production platform, the upstream impact aggregation value, and the downstream impact aggregation value. The preset time window can be a defined time range for abnormal statistics. The abnormal score information can be data related to the abnormal scores of the production platform within the time window (e.g., daily abnormal scores, average scores). The abnormal score information of the target production platform can be the target production platform's own abnormal quantitative data. The upstream impact aggregation value can refer to the sequence of comprehensive abnormal impact values ​​of the upstream neighbor set on the target production platform, calculated using an aggregation function (e.g., weighted average). The downstream impact aggregation value can be the sequence of comprehensive abnormal impact values ​​of the downstream neighbor set on the target production platform, calculated using an aggregation function.

[0058] As an example, firstly, from the set of abnormal score sequences, extract the daily abnormal scores of the target production platform and its upstream and downstream neighboring production platforms within a preset time window (e.g., the last 7 days). Then, for the upstream neighbor set, calculate the weighted sum of their abnormal scores daily based on their coupling strength (edge ​​weight) with the target production platform to obtain the upstream influence aggregation value. Similarly, determine the downstream influence aggregation value. Finally, obtain the abnormal score information of the target production platform, the upstream influence aggregation value, and the downstream influence aggregation value.

[0059] The third step involves fusing the aforementioned target production platform anomaly score information, upstream impact aggregation value information, and downstream impact aggregation value information to obtain comprehensive anomaly information. In practice, firstly, preset weighting coefficients α, β, and γ are assigned to the target production platform anomaly score information, upstream impact aggregation value information, and downstream impact aggregation value information (e.g., α=0.6, β=0.3, γ=0.1). Then, for each day within the time window, after determining the comprehensive anomaly index for that day, the results for each day are arranged in chronological order to form comprehensive anomaly information.

[0060] Step 1052: Based on comprehensive anomaly information, generate supply chain disruption probability information and map it to early warning level information.

[0061] In some embodiments, the aforementioned implementing entity can generate supply chain disruption probability information based on the comprehensive anomaly information, and map it to early warning level information. The supply chain disruption probability information can be a quantitative representation of the likelihood that a target production platform's production will be interrupted due to problems with a key supplier. The early warning level information can refer to discrete level identifiers used to indicate the severity of an event, categorized according to risk level. For example, the early warning level information can include four levels: blue, yellow, orange, and red.

[0062] In some optional implementations of certain embodiments, the aforementioned execution entity may determine the future production platform risk evolution sequence of the target production platform based on the aforementioned comprehensive risk index sequence and using a pre-built time-series prediction model, which may include the following steps: The first step is to construct a corresponding time-series anomaly feature vector based on the aforementioned comprehensive anomaly information. This time-series risk feature vector can be a feature representation extracted from the historical comprehensive risk index sequence of the target production platform and used as input to the time-series prediction model. In practice, firstly, the comprehensive risk index sequence of the target production platform is obtained. Then, the trend slope, volatility variance, risk acceleration, moving average, and extreme value percentage of the comprehensive risk index sequence are extracted as five types of features. Finally, these five types of features are quantified and integrated to construct the time-series risk feature vector.

[0063] The second step involves constructing a neighbor anomaly aggregation feature vector based on the aforementioned industry topology network and anomaly score sequence. This neighbor risk aggregation feature vector can be a vector that integrates upstream and downstream neighbor risks, such as a vector of upstream weighted risk and downstream weighted risk. In practice, firstly, based on the industry topology network, the upstream neighbor set and the aforementioned downstream neighbor set of the target production platform are determined. Then, the neighbor anomaly score sequence is retrieved, and the upstream aggregation risk, downstream aggregation risk, and neighbor risk fluctuations are calculated by weighting them according to coupling strength. Finally, the neighbor risk aggregation feature vector is obtained by integrating these factors.

[0064] The third step involves inputting the aforementioned time-series anomaly feature vector and the aforementioned neighbor anomaly aggregation feature vector into the pre-built time-series prediction model to obtain the future anomaly evolution sequence of the target production platform. This time-series prediction model can be a model that predicts future trends based on historical time-series data. It can be a bidirectional LSTM model incorporating a topological attention mechanism. The time-series prediction model can adopt a three-layer structure: encoding -> fusion -> decoding. The encoding layer takes as input a concatenated matrix of the target production platform's own time-series risk feature vector and the neighbor risk aggregation feature vector, and outputs a hidden state sequence. The fusion layer takes as input the hidden state sequence, and through attention weighting or pooling operations, outputs a fixed-length comprehensive time-series feature vector. The decoding layer takes as input the comprehensive time-series feature vector, and through mapping via a fully connected network, outputs the production platform's risk evolution sequence. In practice, firstly, the time-series risk feature vector and the neighbor risk aggregation feature vector are input into a pre-trained bidirectional LSTM time-series prediction model (including a topological attention mechanism). Then, the future risk evolution sequence of the target production platform is obtained through the output layer of the time-series prediction model.

[0065] The fourth step is to identify the set of key entities based on the aforementioned industry topology network. This set of key entities can be a small number of upstream suppliers that are crucial to the normal operation of the target production platform, selected based on the industry topology network.

[0066] Fifth, for each key object in the key object set, perform the following operations: Sub-step one: Based on the aforementioned anomaly evolution sequence, determine the aggregated anomaly representative value corresponding to the aforementioned key object. This aggregated anomaly representative value can be a comprehensive quantitative value of the future risk of the key supplier; for example, the average risk of a key object over the next 15 days is 0.78. In practice, firstly, an anomaly prediction sequence for the key object's future (e.g., the next 14 days) is extracted from the anomaly evolution sequence. Then, according to a preset aggregation strategy (e.g., taking the maximum value or average value), the anomaly prediction sequence is compressed into a single aggregated anomaly representative value.

[0067] The sixth step is to transform each aggregated anomaly representative value into a supplier disruption probability set. This supplier disruption probability set can be a set of probability values ​​for each supplier in the key supplier set that a supply disruption will occur. In practice, firstly, the aggregated anomaly representative value of each key supplier is directly or through a preset mapping function converted into a corresponding probability value. Then, the probability values ​​of all key suppliers are combined into a probability set.

[0068] Step 7: Based on the aforementioned supplier disruption probability set, determine the supply chain disruption probability to generate supply chain disruption probability information. This disruption probability can be the overall probability of the supply chain breaking due to a key supplier's disruption. In practice, first, obtain the disruption probability of each key supplier from the supplier disruption probability set. Then, based on probability theory, assuming these supplier disruption events are nearly independent, determine the probability that the entire chain will not break (i.e., the probability that all suppliers will not be disrupted). Finally, subtract this probability of no disruption from 1 to obtain the overall probability that at least one key supplier's disruption will lead to the supply chain breaking, which serves as the supply chain disruption probability information.

[0069] The eighth step involves matching the aforementioned supply chain disruption probability information with multiple preset probability threshold intervals to generate an initial warning level. These preset probability threshold intervals can be predefined numerical ranges that map continuous probability values ​​to different warning levels. For example, the threshold intervals could be set as [0, 0.3) -> green, [0.3, 0.5) -> yellow, [0.5, 0.7) -> orange, [0.7, 1.0] -> red. The initial warning level can be a preliminary classification result directly obtained after matching the threshold intervals. For example, if the supply chain disruption probability is 0.65, falling into the orange interval, the initial warning level would be "orange warning".

[0070] Step nine involves generating warning level information based on the initial warning level and the predefined warning level mapping rule base. This predefined rule base can be a knowledge base storing the correspondence between warning levels and more detailed warning parameters (e.g., signal strength, suggested points of concern). For example, in the rule base, "orange warning" might be mapped to: signal strength = 0.7.

[0071] Step 106: Generate corresponding resource scheduling instructions based on the early warning level information.

[0072] In some embodiments, the executing entity may generate corresponding resource scheduling instructions based on the aforementioned warning level information. These resource scheduling instructions may be executable commands that drive resource scheduling interventions.

[0073] In some optional implementations of certain embodiments, the execution entity may generate corresponding resource scheduling instructions based on the aforementioned warning level information, which may include the following steps: The first step involves generating an initial resource scheduling strategy information set based on the aforementioned early warning level information and a pre-defined resource scheduling strategy knowledge base. This pre-defined knowledge base can be a rule or case library storing different recommended resource scheduling schemes. For example, it could include the strategy of "activating backup suppliers and increasing safety stock" when the early warning signal strength is {highest, global}. In practice, firstly, the pre-defined resource scheduling strategy knowledge base is searched based on the early warning signal strength to match eligible resource scheduling strategies. Finally, the eligible resource scheduling strategies are combined into an initial resource scheduling strategy information set.

[0074] The second step involves encoding the initial resource scheduling policy information set into a standard instruction format to generate resource scheduling instructions. In practice, the system first acquires the initial resource scheduling policy information set. Then, based on a predetermined standard instruction format (e.g., a specific JSON structure or API call specification), the policy information is converted into instruction codes or data messages that the resource control system can directly recognize and execute. Finally, standardized resource scheduling instructions are generated.

[0075] Step 107: In response to receiving the resource scheduling instruction, control the resource control terminal to perform resource scheduling operations.

[0076] In some embodiments, the execution entity may, in response to receiving the resource scheduling instruction, control the resource control terminal to perform resource scheduling operations. The resource control terminal may be a hardware or software terminal that performs resource scheduling operations, such as a supplier resource allocation platform.

[0077] In addressing the technical challenges of the aforementioned background technologies, and considering the application scenario—resource scheduling for automakers facing the risk of production stoppages at key component suppliers, especially in extreme cases of "single-source dependence, low safety stock, and extremely short resumption windows"—the following technical problems often arise: Under conditions of heterogeneous resources, complex networks, and dynamically changing states, resource scheduling instructions are difficult to execute reliably, and are prone to failure or execution risks due to resource unavailability or inconsistent states, wasting response time and incurring losses due to supply chain disruptions. To meet the following requirements for this application scenario—instruction robustness, execution determinism, resource adaptability, and process controllability—we have decided to adopt the following solution: In some optional implementations of certain embodiments, the execution entity may, in response to receiving the resource scheduling instruction, control the resource control terminal to perform scheduling operations on the intervention resource, which may include the following steps: The first step is to generate the corresponding resource scheduling operation type and target resource identifier based on the aforementioned resource scheduling instructions. The resource scheduling operation type can be a specific action category for resource scheduling, such as resource allocation or alternative supplier integration. The target resource identifier can be the unique identification information of the resource to be scheduled, for example, "alternative supplier -> B production platform". In practice, the core operation command (i.e., resource scheduling operation type) and operation object (i.e., target resource identifier) ​​can be extracted from the standardized fields of the resource scheduling instructions.

[0078] The second step involves invoking the corresponding resource control interface based on the resource scheduling operation type to establish a communication connection with the target resource control terminal. This resource control interface can be a standardized programming interface provided by the resource control terminal for receiving instructions and exchanging data. The communication connection can be a secure and stable data channel established with the target resource control terminal. In practice, first, based on the resource scheduling operation type, the target system and its corresponding resource control interface to be invoked are determined. Then, a handshake authentication is performed with the target resource control terminal through this interface. Finally, a secure communication connection is established.

[0079] The third step involves sending a resource status query request to the target resource control terminal via the aforementioned communication connection to obtain real-time resource status data. This resource status query request can be an instruction sent to the resource control terminal to retrieve the current status information of a resource. The real-time resource status data can be the latest status information of the resource returned by the resource control terminal at the time of the query (e.g., "Material A, current inventory: 1500 units"). In practice, first, a resource status query request conforming to the target terminal's specifications is constructed through the established communication connection. Then, this request is sent to the target resource control terminal. Finally, the response returned by the terminal is received, and the required real-time resource status data is extracted from it.

[0080] The fourth step involves verifying the current executability of the resource scheduling instruction based on the aforementioned real-time resource status data and target resource identifier, thereby generating an instruction verification result. This verification result can be a Boolean judgment indicating whether the instruction is executable, derived by comparing the instruction requirement with the real-time resource status. For example, if the instruction requires the allocation of 2000 units, but the inventory only contains 1500 units, the verification result would be "unexecutable."

[0081] The fifth step involves generating a resource control command sequence in response to the executable result of the aforementioned instructions. This resource control command sequence can be an ordered set of low-level operation instructions that can be directly understood and executed by the resource control terminal. In practice, firstly, based on the detailed requirements of the resource scheduling instructions, they are decomposed and translated into a series of low-level, atomic operation commands that the resource control terminal can directly understand. Finally, these operation commands are arranged in execution order to form the resource control command sequence.

[0082] The sixth step involves the target resource control terminal receiving the resource control command sequence and executing a scheduling operation on the intervention resource. In practice, firstly, the system sends each command in the resource control command sequence to the target resource control terminal sequentially via a communication connection. Then, upon receiving the command, the resource control terminal drives the physical equipment it controls (e.g., robotic arms, conveyor belts, vehicles) to perform the corresponding actual operation. Finally, the resource in the physical world (e.g., goods) begins to undergo the expected displacement or state change, confirming that the scheduling operation has been executed.

[0083] The above-described operational steps, as an inventive point of this disclosure, solve the technical problem mentioned in the background art: "Under the real-world conditions of heterogeneous resources, complex networks, and dynamically changing states, resource scheduling instructions are difficult to execute reliably, and are prone to instruction failure or execution risks due to resource unavailability or inconsistent states, wasting valuable emergency response time and causing losses due to supply chain disruptions." The reasons for these technical problems are as follows: Existing systems generate scheduling instructions that are mostly human-oriented task lists, lacking mechanisms for interacting with automated terminals, verifying the real-time status of terminals, and generating underlying control instructions. This makes them unable to cope with unexpected situations at the execution site (e.g., equipment failure, network interruption), resulting in 'instruction idling' or uncontrollable execution effects. This invention establishes a closed-loop control link from instruction parsing, connection establishment, status verification to instruction generation and execution, enabling the reliable and intelligent driving of abstract business instructions into specific resource control actions. This saves rework costs and time delays caused by instruction failures, execution errors, or resource conflicts, and improves the accuracy and success rate of resource scheduling.

[0084] In addressing the technical challenges of the aforementioned background technologies, and considering the specific application scenario—the nationwide supply disruption of critical components during automobile production—requiring the rapid allocation of materials from inter-provincial alternative suppliers, which often presents the following technical problems: ambiguous resource scheduling instructions, disordered operation units, and logical conflicts during execution, leading to emergency response failures, increased scheduling response time, wasted scheduling resources, production discontinuity, and production losses. To meet the specific requirements of this application scenario—precise instruction parsing, atomic operation units, ordered execution processes, and automated fault handling—we have decided to adopt the following solution: In some optional implementations of certain embodiments, the execution entity may generate a resource control command sequence in response to the instruction verification result being executable, which may include the following steps: The first step is to perform semantic parsing on the above resource scheduling instructions to extract the corresponding operational objectives and constraints. The operational objective can be the final state that the scheduling instruction aims to achieve. For example, the operational objective could be "Supply 1000 battery cell modules from the Wuhan warehouse to the Shanghai factory within 8 hours." The operational constraints can be the restrictions that must be followed when executing the instruction. For example, the operational constraints could include: "Humidity during transportation <10%", "Anti-static packaging required", and "Full GPS tracking".

[0085] The second step involves decomposing the resource scheduling instructions into a set of basic operation units based on the aforementioned operational objectives and constraints. This set of basic operation units can be the minimum set of executable actions that constitute a complete instruction. For example, the set of basic operation units could include four units: querying inventory, generating a transfer order, scheduling a temperature-controlled vehicle, and arranging receipt.

[0086] The third step involves adding precondition and follow-up effect identifiers to each basic operation unit in the aforementioned set of basic operation units, based on the aforementioned operational constraints, to generate multiple atomic operation tasks. The precondition identifier can indicate a state that must be met before an operation unit can begin. For example, the precondition identifier could be that the precondition for scheduling a temperature-controlled vehicle is "the transfer order has been approved and there are available vehicles." The follow-up effect identifier can indicate the state change resulting from the successful execution of an operation unit. For example, the follow-up effect identifier could be that "outbound scan" results in "inventory reduction, and the goods status changes to 'in transit'." These multiple atomic operation tasks can be multiple independently schedulable basic operation units with accompanying precondition and follow-up effect identifiers. In practice, first, the prerequisites for executing each basic operation unit are determined. Then, the state change of the system after its execution is determined. Finally, precondition and follow-up identifiers are added to each basic operation unit to form multiple atomic operation tasks.

[0087] The fourth step is to construct a task execution dependency graph based on the aforementioned multiple atomic operation tasks. This dependency graph can be a directed acyclic graph (DAG), where nodes represent atomic tasks and edges represent the sequential dependencies between tasks. In practice, firstly, by traversing multiple atomic operation tasks, if a task's precondition is a subsequent effect of another task, a directed edge is created from the precondition task to the first task. Then, a DAG is formed to define the execution order of each task.

[0088] The fifth step involves generating standardized operation instruction templates based on the task execution dependency graph and the pre-built instruction template library. These standardized operation instruction templates can be instruction frameworks that conform to the current task and device interface. In practice, first, each task in the task execution dependency graph is traversed, and a matching template is searched in the instruction template library based on its operation type (e.g., "vehicle dispatching," "temperature control activation"). Then, the most suitable template is selected for each task. Finally, a standardized operation instruction template arranged in task order is output.

[0089] The sixth step involves filling the aforementioned real-time resource status data, target resource identifier, and corresponding scheduling parameters into the standardized operation instruction template to generate device control instructions. These device control instructions can be specific instructions generated by filling specific parameters into the standardized template and can be directly executed by the device.

[0090] Step 7: Based on the timing and logical relationships of the task execution dependency graph, the device control instructions are sorted and synchronized to obtain a preliminary command sequence. This preliminary command sequence can be a list of logically sequentially executable device control instructions obtained by sorting according to the task execution dependency graph. In practice, firstly, the device control instruction set is topologically sorted according to the task execution dependency graph to obtain a basic execution order. Then, concurrently executable task branches in the task execution dependency graph are identified, and concurrent start markers are added to the sequence. Finally, the preliminary command sequence is generated.

[0091] Step 8 involves injecting fault-tolerant processing logic into the initial command sequence to generate a resource control command sequence. This fault-tolerant processing logic can be pre-embedded in the instruction sequence to handle execution anomalies and provide checks and recovery logic. For example, after the "Dispatch Vehicle" command, inject "If no confirmation is received within 10 minutes, attempt to dispatch backup vehicle 002". In practice, first, analyze the potential failure points of each critical instruction in the initial sequence (e.g., dispatch timeout, equipment failure). Then, insert conditional statements, retry instructions, or backup plan instructions after the corresponding instructions. Finally, generate a resource control command sequence with robust error handling logic.

[0092] The above-described operational steps, as an inventive point of this disclosure, solve the technical problem mentioned in the background art: "The resource scheduling instructions are semantically ambiguous and the operation units are disordered, which easily leads to logical conflicts during execution, causing emergency response failures, increasing scheduling response time, wasting a large amount of scheduling resources, resulting in production discontinuity and production losses." The reasons for the above technical problems are as follows: the scheduling instructions are not semantically parsed and atomically decomposed, task dependency relationships are not sorted out, and there is no fault-tolerant processing logic support. This invention achieves standardized transformation of resource scheduling instructions into executable command sequences through instruction semantic parsing -> atomic task decomposition -> dependency relationship modeling -> fault-tolerant logic injection, saving instruction debugging and fault repair time and improving emergency response efficiency.

[0093] The above-described embodiments of this disclosure have the following beneficial effects: The resource scheduling method based on industrial topology networks, as described in some embodiments of this disclosure, achieves automated intervention in supply chain disruptions on production platforms, saving time and resources and reducing losses from supply chain interruptions. Specifically, the causes of supply chain disruption losses are: existing technologies suffer from severe data silos, making it difficult to correlate multi-source heterogeneous raw data, and lack risk propagation analysis from a supply chain perspective, leading to delayed risk identification, insufficient breakpoint prediction, and an inability to take timely intervention measures, resulting in supply chain disruptions. Based on this, the resource scheduling method based on industrial topology networks, as described in some embodiments of this disclosure, firstly, in response to detecting that the inventory of a target product is below a target threshold, generates a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the supply chain corresponding to the target product. The multi-source heterogeneous raw data is uniformly encoded into a standardized time-series vector, forming a continuous and measurable operational information set, providing a unified and comparable basic data representation for subsequent analysis. Then, using an anomaly detection engine, a corresponding anomaly score sequence set is generated based on the above operational information set. By generating the anomaly score sequence set, the degree of deviation of the production platform's state is accurately quantified, improving the sensitivity of early weak anomaly signal identification and providing an intuitive quantitative basis for risk assessment. Next, based on the aforementioned operational information set, the parallel computing node cluster of the topology construction server is invoked to generate an industry topology network representing the product flow relationships between production platforms. By constructing this industry topology network, supply relationships between production platforms can be efficiently mined, supporting real-time monitoring of large-scale production platforms and clearly presenting the industry chain structure. Secondly, the aforementioned anomaly score sequence set and industry topology network are synchronously stored in a shared storage unit for access by the hardware accelerator. This shared storage unit is used for zero-copy data parallel access. This eliminates the data copying overhead between the CPU and the hardware accelerator, ensuring data consistency and providing a low-latency, high-bandwidth data supply channel for subsequent hardware-accelerated computing. Thirdly, through the parallel computing pipeline of the hardware accelerator, the following generation steps are executed: Based on the aforementioned anomaly score sequence set and industry topology network, comprehensive anomaly information for the target production platform is generated. The parallel computing pipeline improves the efficiency of multi-source data fusion, integrates anomaly scores and topology relationships, and achieves a comprehensive assessment of production platform risks, avoiding the one-sidedness of isolated analysis. Based on the aforementioned comprehensive anomaly information, industry chain breakage probability information is generated and mapped to early warning level information. The risk of supply chain disruption is quantified by presenting the risk level intuitively through early warning levels, facilitating rapid response and addressing the delays caused by ambiguous risk levels. Next, based on the aforementioned early warning level information, corresponding resource scheduling instructions are generated. This achieves precise matching between anomalies and scheduling strategies, ensuring the targeted and effective nature of the scheduling instructions. Finally, in response to receiving the aforementioned resource scheduling instructions, the resource control terminal executes resource scheduling operations, including: the storage of components related to the target product and the transportation of these components.Rapidly deploying component storage and transportation scheduling, timely replenishing production resources, reducing the probability of supply chain disruptions, ensuring the continuity of target product production, responding quickly to risks, and minimizing losses from supply chain disruptions.

[0094] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a resource scheduling device based on an industrial topology network. These device embodiments are similar to... Figure 1 Corresponding to the method embodiments shown, this resource scheduling device based on industrial topology networks can be specifically applied to various electronic devices.

[0095] like Figure 2 As shown, a resource scheduling device 200 based on an industrial topology network includes: a first generation unit 201, a second generation unit 202, a scheduling unit 203, a storage unit 204, a calling unit 205, a third generation unit 206, and a control unit 207. The first generation unit 201 is configured to generate a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product, in response to detecting that the inventory of a target product is lower than a target threshold. The second generation unit 202 is configured to generate a corresponding abnormal score sequence set based on the operational information set using an anomaly detection engine. The calling unit 203 is configured to call the parallel computing node cluster of the topology construction server based on the operational information set to generate an industrial topology network representing the product flow relationship between production platforms. The storage unit 204 is configured to synchronously store the abnormal score sequence set and the industrial topology network in a shared storage unit for access by a hardware accelerator, wherein the shared storage unit is used for zero-copy data parallel access. Execution unit 205 is configured to perform the following generation steps through the parallel computing pipeline of the aforementioned hardware accelerator: generating comprehensive anomaly information for the target production platform based on the aforementioned anomaly score sequence and the aforementioned industry topology network; generating supply chain disruption probability information based on the aforementioned comprehensive anomaly information, which is then mapped to early warning level information. Third generation unit 206 is configured to generate corresponding resource scheduling instructions based on the aforementioned early warning level information. Control unit 207 is configured to, in response to receiving the aforementioned resource scheduling instructions, control the resource control terminal to perform resource scheduling operations, wherein the aforementioned scheduling operations include: storage of components related to the target product and transportation of the aforementioned components.

[0096] It is understandable that the units described in the resource scheduling device 200 based on the industrial topology network are related to the reference... Figure 1The steps in the described method correspond accordingly. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the resource scheduling device 200 based on the industrial topology network and the units contained therein, and will not be repeated here.

[0097] The following is for reference. Figure 3 It shows a schematic diagram of the structure of an electronic device (e.g., an electronic device) 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0098] like Figure 3 As shown, the electronic device 300 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0099] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0100] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0101] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0102] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0103] The aforementioned computer-readable medium may be included within the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: in response to detecting that the inventory of a target product is below a target threshold, generate a corresponding operational information set based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product; utilize an anomaly detection engine to generate a corresponding anomaly score sequence set based on the aforementioned operational information set; based on the aforementioned operational information set, invoke a parallel computing node cluster of a topology construction server to generate an industrial topology network characterizing the product flow relationship between production platforms; and synchronously store the aforementioned anomaly score sequence set and the aforementioned industrial topology network in shared storage. The unit is provided for access by the hardware accelerator, wherein the shared storage unit is used for zero-copy data parallel access; the following generation steps are performed through the parallel computing pipeline of the hardware accelerator: generating comprehensive anomaly information of the target production platform based on the anomaly score sequence and the industry topology network; generating supply chain disruption probability information based on the comprehensive anomaly information, which is mapped to early warning level information; generating corresponding resource scheduling instructions based on the early warning level information; and controlling the resource control terminal to perform resource scheduling operations in response to receiving the resource scheduling instructions, wherein the scheduling operations include: storage of components related to the target product and transportation of the components.

[0104] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0105] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0106] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a first generation unit, a second generation unit, a scheduling unit, a storage unit, a calling unit, a third generation unit, and a control unit. The names of these units do not necessarily limit the specific unit itself; for example, the first generation unit may also be described as "a unit that, in response to detecting that the inventory of a target product is below a target threshold, generates a corresponding set of operational information based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product."

[0107] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0108] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A resource scheduling method based on an industrial topology network, comprising: In response to detecting that the inventory of a target product is lower than a target threshold, a corresponding set of operational information is generated based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product. Using the anomaly detection engine, a corresponding set of anomaly score sequences is generated based on the aforementioned set of operational information; Based on the aforementioned set of operational information, the parallel computing node cluster of the topology construction server is invoked to generate an industrial topology network that characterizes the product flow relationship between production platforms. The abnormal score sequence set and the industry topology network are synchronously stored in a shared storage unit for access by the hardware accelerator, wherein the shared storage unit is used for zero-copy data parallel access; The following generation steps are performed through the parallel computing pipeline of the hardware accelerator: Based on the set of abnormal score sequences and the industry topology network, comprehensive abnormal information of the target production platform is generated; Based on the comprehensive anomaly information, supply chain disruption probability information is generated and mapped to early warning level information; Based on the aforementioned warning level information, a corresponding resource scheduling instruction is generated; In response to receiving the resource scheduling instruction, the resource control terminal is controlled to perform a resource scheduling operation, wherein the scheduling operation includes: storage of components related to the target product and transportation of the components.

2. The method according to claim 1, wherein, In response to detecting that the inventory of a target product is lower than a target threshold, a corresponding operational information set is generated based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product, including: For each of the multiple heterogeneous raw data from the various production platforms, perform the following steps: The heterogeneous raw data from multiple sources on the production platform are subjected to heterogeneous data time series isomorphism processing to obtain isomorphic time series data; Based on the homogenized time series data, a standard multidimensional trajectory vector is constructed; The standard multidimensional trajectory vector is subjected to continuity preservation processing to generate corresponding running information; All operational information is integrated into an operational information set.

3. The method according to claim 1, wherein, The step of using an anomaly detection engine to generate a corresponding anomaly score sequence set based on the runtime information set includes: For each piece of operational information in the aforementioned operational information set, the following steps are performed: Using the first node of the anomaly detection engine, a trajectory curvature change sequence corresponding to the running information is generated; Using the second node of the anomaly detection engine, based on the trajectory curvature change sequence, the preset historical normal trajectory benchmark of the production platform itself, and the industry average trajectory benchmark of the same period, a dual-benchmark dynamic discrimination operation is performed to obtain dual-benchmark deviation data. Based on the dual-benchmark deviation data and adaptive weight parameters, an anomaly score sequence is generated, wherein the adaptive weight parameters are dynamically determined by the anomaly detection engine based on the completeness of historical data and the accuracy of historical warnings on the production platform.

4. The method according to claim 1, wherein, The step of invoking a cluster of parallel computing nodes of a topology construction server based on the aforementioned runtime information set to generate an industry topology network representing the product flow relationships between production platforms includes: The first parallel computing node of the parallel computing node cluster is invoked to determine the event-type trajectory similarity between every two production platforms in the running information set, so as to generate an event synchronization matrix. The second parallel computing node of the parallel computing node cluster is invoked to determine the numerical correlation of the continuous trajectories of every two production platforms in the running information set, so as to generate a numerical correlation matrix. The third parallel computing node of the parallel computing node cluster is invoked to determine the time-delay causal relationship between the running information of every two production platforms in the running information set, so as to generate the optimal time-delay matrix; The event synchronization matrix, the numerical correlation matrix, and the optimal time delay matrix are weighted and fused to generate a comprehensive similarity matrix between production platforms. Using the pre-acquired importance scores of production platforms, the comprehensive similarity matrix between production platforms is weighted and corrected to generate a coupling strength matrix between production platforms; The coupling strength matrix between the production platforms is filtered to generate an initial directed network of the industrial chain; Noise filtering is applied to the initial directed network of the industrial chain to generate an industrial topology network.

5. The method according to claim 1, wherein, The process of generating comprehensive anomaly information for the target production platform based on the anomaly score sequence set and the industry topology network includes: Based on the industry topology network, extract the upstream neighbor set and downstream neighbor set of the target production platform; Based on the abnormal score sequence set, the abnormal score information corresponding to the target production platform, the upstream neighbor set, and the downstream neighbor set within a preset time window is determined, so as to obtain the abnormal score information of the target production platform, the upstream influence aggregation value information, and the downstream influence aggregation value information. The abnormal score information of the target production platform, the aggregated value information of the upstream impact, and the aggregated value information of the downstream impact are fused to obtain comprehensive abnormal information.

6. The method according to claim 1, wherein, The step of generating supply chain disruption probability information based on the comprehensive anomaly information, and mapping it to early warning level information, includes: Based on the comprehensive anomaly information, a corresponding time-series anomaly feature vector is constructed; Based on the industry topology network and the anomaly score sequence, a neighbor anomaly aggregation feature vector is constructed; The time-series anomaly feature vector and the neighbor anomaly aggregation feature vector are input into the pre-built time-series prediction model to obtain the future anomaly evolution sequence of the target production platform; Based on the aforementioned industry topology network, a set of key objects is determined; For each key object in the key object set, perform the following operations: Based on the abnormal evolution sequence, determine the aggregated abnormal representative value corresponding to the key object; Each aggregated anomaly representative value is transformed into a supplier disruption probability set; Based on the supplier supply disruption probability set, the supply chain disruption probability is determined to generate supply chain disruption probability information. The probability information of supply chain disruption is matched with multiple preset probability threshold ranges to generate an initial warning level; Based on the initial warning level and the predefined warning level mapping rule base, warning level information is generated.

7. The method according to claim 1, wherein, The step of generating corresponding resource scheduling instructions based on the warning level information includes: Based on the aforementioned warning level information and the preset resource scheduling strategy knowledge base, an initial resource scheduling strategy information set is generated; The initial resource scheduling strategy information set is encoded into a standard instruction format to generate resource scheduling instructions.

8. A resource scheduling device based on an industrial topology network, comprising: The first generation unit is configured to generate a corresponding set of operational information based on multi-source heterogeneous raw data from multiple production platforms in the industrial chain corresponding to the target product in response to detecting that the inventory of the target product is lower than the target threshold. The second generation unit is configured to use an anomaly detection engine to generate a corresponding set of anomaly score sequences based on the set of runtime information. The calling unit is configured to call the parallel computing node cluster of the topology construction server based on the running information set to generate an industry topology network that represents the product flow relationship between production platforms. A storage unit is configured to synchronously store the abnormal score sequence set and the industry topology network to a shared storage unit for access by a hardware accelerator, wherein the shared storage unit is used for zero-copy data parallel access; The execution unit is configured to perform the following generation steps via the parallel computing pipeline of the hardware accelerator: Based on the set of abnormal score sequences and the industry topology network, comprehensive abnormal information of the target production platform is generated; Based on the comprehensive anomaly information, supply chain disruption probability information is generated and mapped to early warning level information; The third generation unit is configured to generate corresponding resource scheduling instructions based on the warning level information; The control unit is configured to, in response to receiving the resource scheduling instruction, control the resource control terminal to perform a resource scheduling operation, wherein the scheduling operation includes: storage of components related to the target product and transportation of the components.

9. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-7.

10. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images